分区域多标准的全参考图像质量评价算法
|
摘要
图像质量评价在图像采集、图像压缩、图像传输等领域有着广泛的应用,一直是图像处理领域的研究热点之一.本文提出了一种模拟人的视觉感知过程中的对不同区域敏感度不同的特点,针对待评图像进行分区域采用不同标准的全参考型图像质量评价算法.该算法首先对图像颜色空间由RGB转换到YIQ,使之更符合人类视觉特性;再对其亮度空间进行数学形态学的膨胀计算预处理,并用边缘检测算子标记出图像中所有的边缘像素点;根据5×5的邻域内是否包含边缘点将图像分为纹理区和平滑区域.针对包含边缘特征的纹理区域的结构参数采用梯度进行描述,参考图像和失真图像在像素点的方差表述像素点失真的敏感性;对于平滑区域的像素点采用对比度作为表征结构信息的变量,并使用基于视觉显著性的综合策略;结合失真和参考图像的视觉显著性矩阵、结构相似性矩阵SCR(x)、色彩饱和度相似性矩阵,可分别得到纹理区和平滑区的图像质量评价分区域结果.取两个分区域结果的均值,可得到最后的全图像质量评价指标SMC-IQA.该算法在CSIQ、TID2008和TID2013等3个通用的图像质量评价数据库上进行了实验.实验结果表明与主流的图像质量评测方法相比较,本文所提出的分区域多标准的全参考图像质量评价算法与主观评价的结果具有更好的一致性,更符合人类视觉系统的特性.
Image quality assessment is widely used in image collection,image compression,and image transmission. It is one of the research hotspots in image processing. This article proposes a full-reference image quality assessment algorithm,which simulates human visual perception with varying sensitivity to different regions. With this method,image color space was transformed from RGB to YIQ for consistency with the human visual system. A morphological dilation method was used during pretreatment,and all edge pixels were marked by edge detection operators. Thereafter,the image was segmented into texture region and smooth region according to whether or not the edge points were included in the 5×5 neighborhood. A gradient value was used to assess the structural parameters of the texture region. The variance in reference image and distorted image at a pixel level was used to assess pixel distortion.For pixels in the smooth region,the contrast value was used to assess the structure features,and a synthesis strategy based on visual salience was adopted. The image quality assessment results can be obtained by combining the visual saliency matrix,structure similarity matrix SCR(x),and color saturation matrix of distortion and reference images.The final image quality assessment index(SMC-IQA)was the mean of the results from two kinds of regions. Experiments were conducted on the CSIQ,TID2008,and TID2013 databases. Compared with state-of-the-art image quality assessment methods,experiment results show that this algorithm is closer to subjective assessment index by the human visual system.
Image quality assessment is widely used in image collection,image compression,and image transmission. It is one of the research hotspots in image processing. This article proposes a full-reference image quality assessment algorithm,which simulates human visual perception with varying sensitivity to different regions. With this method,image color space was transformed from RGB to YIQ for consistency with the human visual system. A morphological dilation method was used during pretreatment,and all edge pixels were marked by edge detection operators. Thereafter,the image was segmented into texture region and smooth region according to whether or not the edge points were included in the 5×5 neighborhood. A gradient value was used to assess the structural parameters of the texture region. The variance in reference image and distorted image at a pixel level was used to assess pixel distortion.For pixels in the smooth region,the contrast value was used to assess the structure features,and a synthesis strategy based on visual salience was adopted. The image quality assessment results can be obtained by combining the visual saliency matrix,structure similarity matrix SCR(x),and color saturation matrix of distortion and reference images.The final image quality assessment index(SMC-IQA)was the mean of the results from two kinds of regions. Experiments were conducted on the CSIQ,TID2008,and TID2013 databases. Compared with state-of-the-art image quality assessment methods,experiment results show that this algorithm is closer to subjective assessment index by the human visual system.
引文
[1]卫津津,金志刚.非主观值训练的盲视频质量评价算法[J].天津大学学报:自然科学与工程技术版,2016,49(6):562-566.Wei Jinjin,Jin Zhigang.Blind video quality assessment strategy without subjective scores training[J].Journal of Tianjin University:Science and Technology,2016,49(6):562-566(in Chinese).
[2]李钊,史再峰,李斌桥,等.基于视觉显著性与对比度特性的图像质量评价[J].南开大学学报:自然科学版,2015,48(6):46-52.Li Zhao,Shi Zaifeng,Li Binqiao,et al.Image quality evaluation based on visual saliency and contrast characteristics[J].Journal of Nankai University:Natural Science Edition,2015,48(6):46-52(in Chinese).
[3]温阳,夏小妹,杨琳.基于视觉注意的全参考彩色图像质量评价方法[J].计算机测量与控制,2017,25(6):279-281.Wen Yang,Xia Xiaomei,Yang Lin.A method of full reference color image quality assessment based on visual attention[J].Journal of Computer Measurement and Control,2017,25(6):279-281(in Chinese).
[4]Wang Z,Bovik A C,Sheikh H R,et al.Image quality assessment:From error visibility to structural similarity[J].IEEE Transactions on Image Processing,2004,13(4):600-612.
[5]Wang Z,Li Q.Information content weighting for perceptual image quality assessment[J].IEEE Transactions on Image Processing,2011,20(5):1185-1198.
[6]Yang Y,Ming J.Image quality assessment based on the space smilarity decomposition model[J].Signal Processing,2016,120(3):797-805.
[7]Shi Z F,Chen K X,Pang K,et al.A perceptual image quality index based on global and double-random window similarity[J].Digital Signal Processing,2017,60(1):277-286.
[8]闫钧华,朱可,张婉怡,等.基于显著性图像边缘的全参考图像质量评价[J].仪器仪表学报,2016,37(9):2140-2148.Yan Junhua,Zhu Ke,Zhang Wanyi,et al.Full reference image quality assessment based on significant image edge[J].Journal of Instrumentation,2016,37(9):2140-2148(in Chinese).
[9]Sheikh H R,Bovik A C.Image information and visual quality[J].IEEE Transactions on Image Processing,2006,15(2):430-444.
[10]Shi Z F,Zhang J P,Cao Q J,et al.Full-reference image quality assessment based on image segmentation with edge feature[J].Signal Processing,2018,145(4):99-105.
[11]Chandler D M,Hemami S S.VSNR:A wavelet-based visual signal-to-noise ratio for natural images[J].IEEETransactions on Image Processing,2007,16(9):2284-2298.
[12]Yuan Y,Guo Q,Lu X.Image quality assessment:Asparse learning way[J].Neurocomputing,2015,159(4):227-241.
[13]Yang C C,Kwok S H.Efficient gamut clipping for color image processing using LHS and YIQ[J].Optical Engineering,2003,42(3):701-711.
[14]宋琼琼,贾振红.基于人眼视觉特性的自适应中值滤波算法[J].光电子·激光,2008,19(1):128-130.Song Qiongqiong,Jia Zhenhong.Adaptive median filter algorithm based Oil human visual system[J].Journal of Optoelectronics·Laser,2008,19(1):128-130(in Chinese).
[15]付伟,顾晓东.基于人眼视觉特性的EZW图像编码改进算法[J].微电子学与计算机,2010,27(3):47-50.Fu Wei,Gu Xiaodong.Improved EZW image encoding algorithm based on HVS[J].Microellectronics&Computer,2010,27(3):47-50(in Chinese).
[16]Larson E C,Chandler D M.Most apparent distortion:Full-reference image quality assessment and the role of strategy[J].Journal of Electronic Imaging,2010,19(1):011006 1-21.
[17]Ponomarenko N,Lukin V,Zelensky A,et al.TID2008-adatabase for evaluation of full-reference visual quality assessment metrics[J].Advances of Modern Radioelectron,2009,10(4):30-45.
[18]Ponomarenko N,Ieremeiev O,Lukin V,et al.Color image database TID2013:Peculiarities and preliminary results[C]//Proceeding of European Workshop on Visual Information Processing.Paris,France,2013:106-111.
[19]VQEG.Final Report from the Video Quality Experts Group on the Validation of Objective Models of Video Quality Assessment[EB/OL].http://www.vqeg.org,2000-03-15.
[2]李钊,史再峰,李斌桥,等.基于视觉显著性与对比度特性的图像质量评价[J].南开大学学报:自然科学版,2015,48(6):46-52.Li Zhao,Shi Zaifeng,Li Binqiao,et al.Image quality evaluation based on visual saliency and contrast characteristics[J].Journal of Nankai University:Natural Science Edition,2015,48(6):46-52(in Chinese).
[3]温阳,夏小妹,杨琳.基于视觉注意的全参考彩色图像质量评价方法[J].计算机测量与控制,2017,25(6):279-281.Wen Yang,Xia Xiaomei,Yang Lin.A method of full reference color image quality assessment based on visual attention[J].Journal of Computer Measurement and Control,2017,25(6):279-281(in Chinese).
[4]Wang Z,Bovik A C,Sheikh H R,et al.Image quality assessment:From error visibility to structural similarity[J].IEEE Transactions on Image Processing,2004,13(4):600-612.
[5]Wang Z,Li Q.Information content weighting for perceptual image quality assessment[J].IEEE Transactions on Image Processing,2011,20(5):1185-1198.
[6]Yang Y,Ming J.Image quality assessment based on the space smilarity decomposition model[J].Signal Processing,2016,120(3):797-805.
[7]Shi Z F,Chen K X,Pang K,et al.A perceptual image quality index based on global and double-random window similarity[J].Digital Signal Processing,2017,60(1):277-286.
[8]闫钧华,朱可,张婉怡,等.基于显著性图像边缘的全参考图像质量评价[J].仪器仪表学报,2016,37(9):2140-2148.Yan Junhua,Zhu Ke,Zhang Wanyi,et al.Full reference image quality assessment based on significant image edge[J].Journal of Instrumentation,2016,37(9):2140-2148(in Chinese).
[9]Sheikh H R,Bovik A C.Image information and visual quality[J].IEEE Transactions on Image Processing,2006,15(2):430-444.
[10]Shi Z F,Zhang J P,Cao Q J,et al.Full-reference image quality assessment based on image segmentation with edge feature[J].Signal Processing,2018,145(4):99-105.
[11]Chandler D M,Hemami S S.VSNR:A wavelet-based visual signal-to-noise ratio for natural images[J].IEEETransactions on Image Processing,2007,16(9):2284-2298.
[12]Yuan Y,Guo Q,Lu X.Image quality assessment:Asparse learning way[J].Neurocomputing,2015,159(4):227-241.
[13]Yang C C,Kwok S H.Efficient gamut clipping for color image processing using LHS and YIQ[J].Optical Engineering,2003,42(3):701-711.
[14]宋琼琼,贾振红.基于人眼视觉特性的自适应中值滤波算法[J].光电子·激光,2008,19(1):128-130.Song Qiongqiong,Jia Zhenhong.Adaptive median filter algorithm based Oil human visual system[J].Journal of Optoelectronics·Laser,2008,19(1):128-130(in Chinese).
[15]付伟,顾晓东.基于人眼视觉特性的EZW图像编码改进算法[J].微电子学与计算机,2010,27(3):47-50.Fu Wei,Gu Xiaodong.Improved EZW image encoding algorithm based on HVS[J].Microellectronics&Computer,2010,27(3):47-50(in Chinese).
[16]Larson E C,Chandler D M.Most apparent distortion:Full-reference image quality assessment and the role of strategy[J].Journal of Electronic Imaging,2010,19(1):011006 1-21.
[17]Ponomarenko N,Lukin V,Zelensky A,et al.TID2008-adatabase for evaluation of full-reference visual quality assessment metrics[J].Advances of Modern Radioelectron,2009,10(4):30-45.
[18]Ponomarenko N,Ieremeiev O,Lukin V,et al.Color image database TID2013:Peculiarities and preliminary results[C]//Proceeding of European Workshop on Visual Information Processing.Paris,France,2013:106-111.
[19]VQEG.Final Report from the Video Quality Experts Group on the Validation of Objective Models of Video Quality Assessment[EB/OL].http://www.vqeg.org,2000-03-15.
© 2004-2018 中国地质图书馆版权所有 京ICP备05064691号 京公网安备11010802017129号 地址：北京市海淀区学院路29号 邮编：100083 电话：办公室：(+86 10)66554848；文献借阅、咨询服务、科技查新：66554700 |